Extract Transform Load
ETL stands for Extract Transform and Load.
OLTP (Online Transaction Processing system) generate lots of transactional data . But the OLTP system cannot store these transactional data for so long because of performance issues. So these transactional data are stored in the Data ware house.
Example for OLTP Transaction Data :
Transaction Details of each customer in the case of banking.
The transactional data generated in the different OLTP are stored at multiple locations.They might be in the form of a database, flat ile or XML format etc. Their Schema might be different.. But the Data Warehouse has a unique schema.
First the transactional data must be extracted from each OLTP system.The extracted data is transformed in to a structure to match with that of the Data Warehouse. This transformation is some times termed as data Cleansing.
Some Commonly used Transformations
1. Convert to match with the Data Warehouse's Schema.
2. Duplicate records are removed.
3. Noisy (incomplete / unknown) data are handled according to the business rules.
4. Aggregates are computed if required.